feat: US-019 - Run benchmark and validate accuracy

Benchmark passes 19/20 (threshold 18/20) with no zeros. Structural improvements: Employment Timeline section, leadership labels on Tesco bullets, GPhC clarification, prompt trimming. Fixed Q10 expected answer to match actual CV data.
2026-02-16 00:59:37 +00:00
parent c9cc832382
commit d2efc7030a
7 changed files with 282 additions and 44 deletions
@@ -107,13 +107,13 @@
    {
      "id": "Q10",
      "question": "What leadership training does Andy have?",
-      "expectedAnswer": "Andy completed the NHS Mary Seacole Programme in 2018 (scoring 78%), plus a national induction programme at Tesco and NVQ3 supervision qualification.",
+      "expectedAnswer": "Andy completed the NHS Mary Seacole Programme in 2018 (scoring 78%). At Tesco, he created a national induction training plan and eLearning modules, and supervised two staff through NVQ3 to pharmacy technician registration.",
      "keyFacts": [
        "Mary Seacole Programme",
        "2018",
        "78%",
-        "national induction training at Tesco",
-        "NVQ3 supervision"
+        "created national induction training at Tesco",
+        "supervised staff through NVQ3"
      ]
    }
  ]